Data integration is the problem of combining data residing at different sources, and providing the user with a unified view of these data. It is characterized by an architecture based on a global schema, with the set of integrity constraints, and a set of sources. In this paper we investigate the way in which Closed World Assumption on a source data base can be coherently propagated to the global schema. The problem to resolve is directly connected by the fact that a global schema has a number (possibly infinite) of minimal models, caused by the incompleteness of source databases w.r.t. the integrity constraints over global schema. The aim of this preliminary work is to open the perspective for query language with negation in Data integration framework.